| Wednesday, May 9th, 2012, 00:00 UTC | ||
| [00:00:10] | wagnerrp: | http://pastebin.com/275ztjdZ |
| [00:00:16] | wagnerrp: | output from pre-sync for comparison |
| [00:00:42] | Beirdo: | interesting |
| [00:02:46] | wagnerrp: | SUCCESS! |
| [00:02:48] | wagnerrp: | ... ish |
| [00:03:58] | wagnerrp: | VDPAU plays it |
| [00:04:10] | wagnerrp: | but it seems it just discards any data >256 as garbage |
| [00:04:20] | wagnerrp: | meaning those blocks become garbage blocks |
| [00:04:33] | Beirdo: | hehehe |
| [00:04:34] | wagnerrp: | and persist until a scene change flushes them |
| [00:04:39] | Beirdo: | nasty |
| [00:05:48] | Captain_Murdoch: | jya, MDM downloads are concurrent. I haven't looked at the source, but it sounds like the number of concurrent downnloads may be different per protocol. at least that's what this URL indicates. http://www.qtcentre.org/threads/37530-QNetwor . . . for-one-host |
| [00:06:06] | Captain_Murdoch: | MDM just queues up downloads with QNetworkAccessManager so we're relying on Qt's limites, not any limit in MDM. |
| [00:06:47] | Beirdo: | you could run more than one QNetworkAccessManager if you need more, I would presume |
| [00:07:48] | wagnerrp: | VDPAU, normal/Xv, and normal/OpenGL all play exactly the same |
| [00:08:09] | Beirdo: | yeah, I think we hardcode the pixel format to yuv420p |
| [00:08:12] | wagnerrp: | plays fine, no jerkiness, but artifacts all over the place anywhere there was content >255 |
| [00:08:21] | Beirdo: | which would be incorrect for those files |
| [00:12:24] | Beirdo: | anywho, I think it's time to start heading home before I break something |
| [00:12:52] | jya_: | Captain_Murdoch: to be perfectly clear… I use the MythDownloadManager singleton (as returned by GetMythDownloadManager()) ; I call it from different threads so if thread 1 download file1, thread2 download file2 ; file2 could be received before file1 (whichever is the fastest / shortest), and there's no queue as such ? |
| [00:14:26] | Captain_Murdoch: | there is a queue inside MDM. there's a MDM run thread which pops items off the queue and passes them to either QNetworkAccessManager for non-myth:// downloads or to an internal download thread for myth:// downloads. |
| [00:15:15] | Captain_Murdoch: | the MDM queue is processed as fast as items come in, so your example of #2 finishing before #1 could happen. |
| [00:15:15] | ** MythLogBot http://code.mythtv.org/trac/ticket/2 ** | |
| [00:15:15] | ** MythLogBot http://code.mythtv.org/trac/ticket/1 ** | |
| [00:15:22] | jya_: | sorry, but that's not obvious to me in what the queue does nor how will QNetworkAccessManager will process it.. |
| [00:15:57] | jya_: | ok… that's all I wanted to know… That file2 download doesn't wait for file1 to finish before it starts |
| [00:17:08] | Captain_Murdoch: | the queue is just so you can queue items from one thread and have them all passed to QNetworkAccessManager from a single thread since QNetworkAccessManager is not thread safe. |
| [00:17:27] | Captain_Murdoch: | s/to QNetworkAccessManager/to a single QNetworkAccessManager/ |
| [00:17:51] | jya_: | well, as long as i don't have to wait for a download to finish for another one to start, that's fine |
| [00:19:12] | Captain_Murdoch: | qnetworkaccessmanager doesn't know about the MDM queue, it also has it's own internal queue I assume but I haven't tested what happens if we hit more than 6 HTTP downloads. I just asssume the rest are queued and that QNetworkAccessManager::get/post() don't block. |
| [00:19:18] | Captain_Murdoch: | yeah, no waiting. |
| [00:20:09] | Captain_Murdoch: | and we have multiple threads all using the same QNetworkAccessManager since MDM abstracts access to QNAM. for instance image loading of http:// images on the channel icon download screen. |
| [00:20:35] | Captain_Murdoch: | anyway, more info than you wanted I guess, but might be helpful if you use it later for other things. :) |
| [00:21:14] | jya_: | the way I'm using it, is using the blocking ::download() MDM call; each in a different thread |
| [00:21:33] | jya_: | all using the MDM common singleton |
| [00:21:55] | Captain_Murdoch: | that's fine, the blocking is just done internally in MDM in a while loop waiting for the background download to complete. |
| [00:22:44] | Captain_Murdoch: | it was provided as a convenience to places that already were downloading in their own thread so they didn't have to handle the status/completion events if they downloaded in the background. |
| [00:23:11] | jya_: | yeah, I certainly couldn't be bothered having to deal with signals |
| [00:23:32] | jya_: | to tell another thread that download is complete, I use a mutex |
| [00:23:35] | Captain_Murdoch: | these are myth events, so you'd have to have a ::customEvent() handler. |
| [00:55:10] | skd5aner (skd5aner!~skd5aner@cpe-071-071-242-134.carolina.res.rr.com) has quit (Ping timeout: 276 seconds) | |
| [00:58:25] | Beirdo: | wagnerrp: here's what I get: |
| [00:58:32] | Beirdo: | 2012-05–08 17:57:42.098829 E VidOutVDPAU: Picture format is not supported. |
| [01:00:29] | k-man (k-man!~jason@unaffiliated/k-man) has joined #mythtv | |
| [01:01:23] | wagnerrp: | want logs? |
| [01:03:06] | Beirdo: | if ya want. It flatly refuses to play it post-sync though :) |
| [01:04:13] | [TheAsp] ([TheAsp]!~TheAsp@blk-222-131-25.eastlink.ca) has quit (Quit: ZNC - http://znc.in) | |
| [01:08:29] | Beirdo: | 2012-05–08 18:08:09.613854 I Last hardware profile update was > 30 days ago, update required... |
| [01:08:36] | Beirdo: | that's three days in a row now |
| [01:08:44] | Beirdo: | I think we have a bug |
| [01:08:49] | dmfrey (dmfrey!~dmfrey@64-121-93-243.c3-0.smt-ubr1.atw-smt.pa.cable.rcn.com) has joined #mythtv | |
| [01:22:49] | Dj_FlyBy{MS} (Dj_FlyBy{MS}!~djflyby-m@184.175.18.218) has joined #mythtv | |
| [01:23:53] | Dj_FlyBy{MS}: | trying to install MythTV but I get an error about qt-x11 as I am using another program that requires qt47-x11.. Is there a way to encourage MythTV to use qt47-x11 instead of qt-x11 ? |
| [01:29:10] | sphery: | Dj_FlyBy{MS}: /topic (you want #mythtv-users ) |
| [01:29:16] | rhpot1991: | still trying to figure out a good solution to the mythfilldatabase innodb ext4 journaling issue, thoughts on having mythfilldatabase use a memory table? http://dev.mysql.com/doc/refman/5.5/en/memory . . . -engine.html |
| [01:30:36] | Dj_FlyBy{MS}: | whoooops, sorry I didn't even notice |
| [01:31:12] | Dj_FlyBy{MS} (Dj_FlyBy{MS}!~djflyby-m@184.175.18.218) has left #mythtv ("Leaving") | |
| [01:33:00] | superm1: | it is a temp table after all. at least to me that sounds like it fits well with the purpose of the table better than myiasm or innodb |
| [01:33:10] | Beirdo: | OK. |
| [01:33:53] | Beirdo: | according to nvidia forums, vdpau does *not* support 10bpc as the hardware does not support it as best as the person replying can tell |
| [01:34:17] | Beirdo: | so... morons with 9/10bit video will need to have some processor balls. |
| [01:34:25] | Beirdo: | err, I mean people :) |
| [01:34:54] | Beirdo: | seems it's mainly used by anime types, recoding 8bit source to 10bit to get better compression |
| [01:35:52] | ** xris wonders what to do about lyngsat and channel icons. ** | |
| [01:37:03] | sphery: | rhpot1991: FWIW, memory based tables for mfdb would be huge--especially for --dd-grab-all --since all the varchars would be padded out to max |
| [01:37:27] | sphery: | rhpot1991: I will likely change the code to request myisam, for now, but still have to do some testing |
| [01:37:59] | rhpot1991: | sphery: I can confirm that using myisam "fixes" the issue |
| [01:38:00] | sphery: | we've changed all the DB schema creation to specify the engine as of 0.25, just have the temp tables for mfdb and scheduler left |
| [01:38:20] | rhpot1991: | but setting that as a default for the whole mysql instance seems like a bad idea |
| [01:38:24] | sphery: | rhpot1991: it would be easier for me to actually get others to not hate the change if you gave me at least a pointer to "the issue" |
| [01:39:26] | rhpot1991: | sphery: would an ubuntu bug be good enough? |
| [01:39:38] | sphery: | yeah |
| [01:39:45] | rhpot1991: | I'll get one for you later tonight |
| [01:39:46] | sphery: | I just need more than "mythfilldatabase innodb ext4 journaling issue" |
| [01:39:48] | superm1: | there's a few places it's shown up, forums too http://ubuntuforums.org/showthread.php?t=1974335 |
| [01:39:55] | sphery: | thanks |
| [01:40:03] | rhpot1991: | I think all we have are forum posts so far |
| [01:40:05] | tgm4883: | sphery, it doesn't appear to be only ext4 |
| [01:40:14] | tgm4883: | although I only have one person saying that |
| [01:40:17] | rhpot1991: | but I've dug into it quite a bit, so I can get a decently detailed bug report for you |
| [01:40:23] | tgm4883: | someone said it happens on XFS as well |
| [01:40:55] | sphery: | all I'm going on is the comments rhpot1991 has made in here (today and May 2)--I haven't seen any other mention of it |
| [01:41:05] | sphery: | this is why I was hoping for a better description |
| [01:41:18] | tgm4883: | that thread has lots of the info |
| [01:41:54] | superm1: | there's another thread with more info i think too |
| [01:42:03] | rhpot1991: | tgm4883: superm1 confirm we don't have a bug already that you know about? |
| [01:42:25] | tgm4883: | rhpot1991, I don't know of a bug |
| [01:43:07] | superm1: | no i dont know of a bug |
| [01:43:15] | sphery: | it definitely would be nice if mysql allowed setting a per-database default (as it does for charset)--would allow you to work around it in the interim--but since it seems you can only set a global or session value for default, that requires changing mysqld config or mythtv code changes |
| [01:43:27] | jya_ (jya_!~jyavenard@mythtv/developer/jya) has quit (Quit: jya_) | |
| [01:43:32] | NightMonkey (NightMonkey!~NightrMon@pdpc/supporter/professional/nightmonkey) has quit (Quit: Body blow! Body blow!) | |
| [01:43:48] | superm1: | http://ubuntuforums.org/showthread.php?t=1956370 is the other forum post people started yelling in |
| [01:44:44] | superm1: | rhpot1991: it would probably be good to get a bug for this filed though anyway, we might end up wanting to SRU something like this too |
| [01:44:49] | sphery: | superm1: just to make sure, the first "thread" is just one post, right? |
| [01:45:00] | tgm4883: | sphery, yep |
| [01:45:09] | sphery: | ok, thanks |
| [01:54:52] | kormoc (kormoc!~kormoc@mythtv/developer/kormoc) has quit (Ping timeout: 272 seconds) | |
| [02:08:36] | tgm4883 (tgm4883!~tgm4883@ubuntu/member/tgm4883) has quit (Excess Flood) | |
| [02:09:20] | stichnot (stichnot!~chatzilla@mythtv/developer/stichnot) has quit (Ping timeout: 256 seconds) | |
| [02:23:45] | stichnot (stichnot!~chatzilla@192.55.54.40) has joined #mythtv | |
| [02:23:46] | stichnot (stichnot!~chatzilla@192.55.54.40) has quit (Changing host) | |
| [02:23:46] | stichnot (stichnot!~chatzilla@mythtv/developer/stichnot) has joined #mythtv | |
| [02:29:32] | tgm4883 (tgm4883!~tgm4883@2001:4968:202:3:20f:eaff:fefc:ba0e) has joined #mythtv | |
| [02:29:32] | tgm4883 (tgm4883!~tgm4883@2001:4968:202:3:20f:eaff:fefc:ba0e) has quit (Changing host) | |
| [02:29:32] | tgm4883 (tgm4883!~tgm4883@ubuntu/member/tgm4883) has joined #mythtv | |
| [02:29:50] | tgm4883 (tgm4883!~tgm4883@ubuntu/member/tgm4883) has quit (Excess Flood) | |
| [02:31:48] | danielk22: | sphery: superm1: It doesn't sound right that InnoDB would be that much slower than MyISAM, 70 minutes vs 6 minutes. And I thought we explicitly tagged all our tables as MyISAM except for some MythWeather ones anyway.. |
| [02:32:37] | sphery: | danielk22: if you read the threads they linked, the problem is that ubuntu configures an innodb data file path in /var/lib/mysql and when we don't specify myisam and they're using mysql 5.5, they get innodb |
| [02:32:55] | sphery: | because /var/lib/mysql is on the root file system which is ext4 + barrier, they get terrible performance |
| [02:33:17] | danielk22: | So our temp tables just use whatever the default engine is? |
| [02:33:28] | sphery: | switching to myisam means that mysql can use the temp file system (which in their case would be tmpfs), so they get fast writes |
| [02:33:57] | danielk22: | Yeah ext4 or xfs + barriers == terrible performance, no matter the DB engine. |
| [02:34:08] | danielk22: | got it |
| [02:34:14] | sphery: | yes, I changed dbcheck.cpp to specify engine (starting with 0.25) for real tables, but didn't change any of the temporary table creation for mfdb or scheduler to use myisam |
| [02:34:55] | danielk22: | At least that sounds like something we can deal with. |
| [02:35:05] | sphery: | so, I'm probably just going to change the temp table creation statements to specify myisam, and maybe even put in a SET storage_engine=MYISAM; for the session (along with SET SESSION sql_mode='' to fix users running with SQL strict mode enabled) |
| [02:36:21] | sphery: | yeah, it seems to be the lesser of multiple evils at this point |
| [02:36:25] | danielk22: | I am wondering which distro will be the first one to drop default ext4 support and turn off barriers by default on ext4 + xfs... |
| [02:36:51] | danielk22: | aka "my next distro" :) |
| [02:37:09] | sphery: | hehe, yeah--with things like firefox "awesome bar" and such, I'd think one will have to |
| [02:38:03] | sphery: | funny how important it is to care about barriers, now, even though we were all happy in our ignorance/ignoring the issue before |
| [02:38:06] | danielk22: | It was firefox performance that lead me to first discover the barrier bug. The impact on MythTV is minimal in comparison. |
| [02:38:40] | danielk22: | sphery: Barriers were turned on to deal with a design bug in ext4. |
| [02:38:50] | sphery: | yeah, awesome bar performance is bad enough for me on ext3 without barriers that I clear history all the time--I can only imagine ext4+barriers |
| [02:38:50] | stuartm (stuartm!~stuartm@mythtv/developer/stuartm) has quit (Remote host closed the connection) | |
| [02:40:19] | danielk22: | Other file-systems have dealt with the power loss issue without corrupting metadata.. XFS has the same delayed allocation that ext4 does, but doesn't have the bug that affects ext4. |
| [02:41:49] | danielk22: | XFS will lose data on power loss that non-delayed allocation filesystems like ext3 won't lose. But only files opened for reading on power-loss are endangered, with EXT4 you can lose everything. |
| [02:42:15] | tgm4883` (tgm4883`!~tgm4883@2001:4968:202:3:20f:eaff:fefc:ba0e) has joined #mythtv | |
| [02:47:32] | tgm4883` is now known as tgm4883 | |
| [02:47:53] | tgm4883 (tgm4883!~tgm4883@2001:4968:202:3:20f:eaff:fefc:ba0e) has quit (Changing host) | |
| [02:47:54] | tgm4883 (tgm4883!~tgm4883@ubuntu/member/tgm4883) has joined #mythtv | |
| [03:12:52] | jpabq: | Beirdo, current master won't build under Fedora16 (crystalhd.c unknown type name HANDLE, etc). I may have time tomorrow night to hack on it, if you need me to. Personally, I am not in a hurry, so no pressure. |
| [03:13:36] | Beirdo: | interesting |
| [03:14:01] | Beirdo: | K, let me know if I can help, or get me some build log output, I guess |
| [03:14:12] | Beirdo: | or I can install FC16 in a VM at work too :) |
| [03:14:13] | Beirdo: | hehe |
| [03:15:53] | jpabq: | I don't have a crystalhd, so I could probably just disable that if I needed to. But, like I said I am not currently concerned about it. I will email you a compile log just in case you feel inclined. |
| [03:16:44] | Beirdo: | sure, cool |
| [03:18:44] | jpabq: | I bet it just needs some include added to the top. |
| [03:21:27] | dmfrey (dmfrey!~dmfrey@64-121-93-243.c3-0.smt-ubr1.atw-smt.pa.cable.rcn.com) has quit (Quit: Ex-Chat) | |
| [03:23:17] | Beirdo: | probably |
| [03:39:50] | jya: | Beirdo: something that may be relevant to you |
| [03:41:10] | jya: | when I fixed the VAAPI detection in configure (it was all screwed up) ; I noticed that the detection for crystalhd was just as bad (same author). However, I didn't touch it… my guess is its just the autodection that isn't working properly, and cyrstalhd compilation is activated when it shouldn't |
| [03:41:31] | Beirdo: | that's quite likely :) |
| [03:41:59] | jya: | I can have a look later, I left it untouched because I had no way of testing it |
| [03:42:32] | jya: | if you check my VAAPI detection commit, you can see what I did.. the crystalhd stuff is just above in the configure |
| [03:42:36] | Beirdo: | k. You might wanna make sure that your VAAPI changes stuck too |
| [03:42:39] | Beirdo: | k |
| [03:43:09] | Beirdo: | right now, I'm wrestling with the Vista VM to move its page file off iSCSI and onto a local drive |
| [03:43:52] | Beirdo: | I so hate computers some days :) |
| [03:44:37] | jya: | wy don't you just disable paging for the purpose of testing ? |
| [03:44:47] | jya: | why would you bother with it? |
| [03:45:15] | Beirdo: | because compiling takes random amounts of memory |
| [03:45:31] | Beirdo: | well not truly random, but unpredictable |
| [03:45:50] | jya: | just allocate more :) |
| [03:47:11] | Beirdo: | I could, but... Windows likes having a paging file regardless |
| [03:47:30] | jya: | it can be disabled. |
| [03:47:52] | jya: | Start Program -> Right click on Computer -> Properties. in the advanced option , you can turn it off |
| [03:47:55] | Beirdo: | yeah, I know |
| [04:00:53] | jpabq_ (jpabq_!~jpabq@71-37-152-118.albq.qwest.net) has joined #mythtv | |
| [04:04:26] | Beirdo: | OK, we'll see if that makes it less crash-prone |
| [04:05:58] | jpabq_ (jpabq_!~jpabq@71-37-152-118.albq.qwest.net) has quit (Ping timeout: 272 seconds) | |
| [04:27:49] | kurre2 (kurre2!~tomimo@xdsl-83-150-88-111.nebulazone.fi) has joined #mythtv | |
| [04:28:01] | jya: | Beirdo: didn't you split the mpegts code from ffmpeg this time ? |
| [04:28:08] | Beirdo: | yes |
| [04:28:16] | Beirdo: | after merging |
| [04:28:36] | jya: | so the amount of patches should be at a minimum now right? |
| [04:28:44] | Beirdo: | heh, hardly |
| [04:29:02] | jya: | that's there's a few, doesn't mean it's not a "minimum" :) |
| [04:29:13] | Beirdo: | well, yeah, I guess |
| [04:30:36] | jya: | in the mean time, iOS 5.1.1 broke something with airplay+raop… sigh |
| [04:31:40] | jya: | they start raop for audio only, which open the audio device…. then start airplay, which also opens the audio device… when playback has started, they close raop. problem is for airplay, opening the audio device failed as it's in use by raop |
| [04:32:19] | jya: | i'm guessing they did so, so if playing a http live stream, you still have audio playing while frontend is seeking for its own content |
| [04:32:57] | amessina (amessina!~amessina@2001:470:1f11:a4:d6be:d9ff:fe8d:7c1e) has joined #mythtv | |
| [04:33:02] | jya: | all fine under iOS or Mac where all audio device allows you to mux ; not fine with alsa digital device not allowing to be opened more than once at a time |
| [04:33:47] | jya: | i would need to merge raop + airplay as a one and only service, so airplay is aware of what raop is doing... |
| [04:39:09] | Beirdo: | have fun :) |
| [04:39:16] | jpabq_ (jpabq_!~jpabq@71-37-152-118.albq.qwest.net) has joined #mythtv | |
| [04:47:42] | jpabq_ (jpabq_!~jpabq@71-37-152-118.albq.qwest.net) has quit (Quit: Colloquy for iPhone - http://colloquy.mobi) | |
| [04:51:53] | Beirdo: | wagnerrp: you up? |
| [04:52:19] | wagnerrp: | aye |
| [04:52:50] | Beirdo: | you know much about the internals of smolt? |
| [04:53:15] | Beirdo: | I'm getting a 500 back from the /client/add_json post |
| [04:54:14] | Beirdo: | just wondering if it's half-borked, etc, and how we'd know |
| [04:54:15] | Beirdo: | heh |
| [04:54:47] | wagnerrp: | i just do the internals, not the protocol |
| [04:54:50] | wagnerrp: | let me check... |
| [04:54:57] | Beirdo: | ah, k. |
| [04:56:00] | Beirdo: | I see the stuff is all owned by group 104 (numeric) |
| [04:56:43] | wagnerrp: | looks owned by 493 to me |
| [04:56:58] | Beirdo: | in /opt/smolt? |
| [04:57:07] | wagnerrp: | smoon |
| [04:57:57] | Beirdo: | k |
| [04:58:04] | Beirdo: | so just the logs go in /opt/smolt |
| [04:58:05] | Beirdo: | heh |
| [04:58:26] | Beirdo: | web logs tht is |
| [04:58:31] | wagnerrp: | looks like its been sending out 500s since about noon today |
| [04:58:45] | wagnerrp: | logs go in smoon to |
| [04:58:46] | wagnerrp: | o |
| [04:59:12] | Beirdo: | yeah, the ones from apache (for the rewrite) are in /opt/smolt, and its internals are in smoon |
| [04:59:17] | Beirdo: | I get it now :) |
| [05:00:32] | wagnerrp: | sheesh... 1060 users without video sources |
| [05:00:48] | Beirdo: | well that would be why my frontend keeps updating... |
| [05:00:51] | Beirdo: | or trying to |
| [05:01:21] | wagnerrp: | only since about noon your time |
| [05:01:25] | Beirdo: | or 1060 users where the scripts lied? :) |
| [05:01:27] | wagnerrp: | prior to that, it was sending 200s |
| [05:01:30] | Beirdo: | did it yesterday too |
| [05:02:32] | Beirdo: | grep 71.212.59.60 server.log |
| [05:02:35] | Beirdo: | that's me |
| [05:02:43] | wagnerrp: | seems to be intermittent actually |
| [05:03:28] | Beirdo: | sucktacular |
| [05:03:36] | wagnerrp: | the scripts dont lie |
| [05:03:43] | Beirdo: | but at least I don't need to worry about it being on my end, it seems |
| [05:03:49] | wagnerrp: | counts all sources in the database with an input mapped to it |
| [05:03:54] | Beirdo: | nice |
| [05:04:22] | wagnerrp: | that really is 43.6% of users submitting a profile from a machine that could not possibly work |
| [05:04:47] | Beirdo: | probably by running the frontend the first time before setup |
| [05:05:08] | Beirdo: | I'm sure it will update in a month (assuming the server doesn't refuse like it is for me) |
| [05:06:19] | Beirdo: | but I guess we'll see |
| [05:06:48] | Beirdo: | seems odd that there'd be that many if it weren't something systematic like that |
| [05:11:14] | wagnerrp: | running 2.4GB virt, seems awfully high |
| [05:11:29] | wagnerrp: | i could restart it |
| [05:11:31] | Beirdo: | unless it's Java :) |
| [05:11:41] | Beirdo: | might be worth a try |
| [05:13:36] | wagnerrp: | the init script, it does nothing |
| [05:13:50] | Beirdo: | lovely |
| [05:15:25] | wagnerrp: | try again |
| [05:16:25] | Beirdo: | 500 |
| [05:19:15] | wagnerrp: | worked for me |
| [05:19:41] | Beirdo: | well, I dunno |
| [05:19:54] | wagnerrp: | retry failed |
| [05:20:24] | Beirdo: | it's consistently denying me. I guess we'll have to set Mr. jams on it |
| [05:41:22] | jya: | Beirdo: in regards to configure vs ffmeg configure, you should have a look at what Torc has done, they did just that |
| [05:48:10] | superm1: | do you guys think you'll relax the libx264 requirement to a lesser version, or is that the hard requirement? If it's the hard requirement, i think i'll just backport newer x264 to the 0.26 PPA |
| [05:48:15] | superm1: | (to older ubuntu's) |
| [05:48:54] | jya: | superm1: that would be a ffmpeg requirements |
| [05:48:59] | jya: | not a myth one |
| [05:49:35] | superm1: | ah so it's probably a real requirement then if ffmpeg is introducing it |
| [05:49:51] | jya: | it came with the ffmpeg resync |
| [05:50:02] | jya: | we don't use libx264 directly, only through ffmpeg |
| [05:50:47] | superm1: | right, so since they're insisting on that newer version, they probably had good reason to bump it up in the first palce |
| [05:51:12] | jya: | probably |
| [05:51:22] | jya: | i'm surprised the bump is only now |
| [05:51:34] | jya: | mplayer pretty require a new version of libx264 every month or so |
| [05:51:55] | superm1: | well the ffmpeg version before this was over a year old wasn't it? |
| [05:53:16] | jya: | still… ffmpeg doesn't impose a change of libx264 that often, while mplayer does |
| [05:53:29] | jya: | it's quite easy to package lib264 anyway |
| [05:53:55] | jya: | and best, you can have two versions installed at once |
| [05:54:01] | jya: | different soname |
| [05:54:24] | jya: | so installing a new one, never breaks an existing install |
| [05:54:32] | jya: | so long as you keep both |
| [05:59:05] | dekarl1 (dekarl1!~dekarl@p4FE84F77.dip.t-dialin.net) has joined #mythtv | |
| [06:00:24] | dekarl (dekarl!~dekarl@p4FE844B1.dip.t-dialin.net) has quit (Ping timeout: 265 seconds) | |
| [06:00:31] | superm1: | i'll see what's doable when i take a look |
| [06:01:00] | jya: | if you check my ubuntu repo |
| [06:01:06] | jya: | I haven't updated in a while |
| [06:01:24] | jya: | but the source package has everything to build it, and upgrade the soname accordingly |
| [06:03:33] | superm1: | cool |
| [06:07:33] | Beirdo: | well, we only sync once per release |
| [06:07:50] | Beirdo: | so yeah, they likely have changed many times |
| [06:15:08] | Beirdo: | I have like 4 versions of libx264 on my boxes, but you can only have one -dev package the way the stock ubuntu packages are |
| [06:15:48] | Beirdo: | running lucid with libx264–120 from precise (currently) |
| [06:24:26] | Beirdo: | anyways... bed |
| [06:34:16] | stichnot (stichnot!~chatzilla@mythtv/developer/stichnot) has quit (Remote host closed the connection) | |
| [06:41:45] | Steve_Goodey (Steve_Goodey!~steve@host86-182-4-91.range86-182.btcentralplus.com) has joined #mythtv | |
| [06:45:12] | solars (solars!~solars@clnet-kmu02-090.ikbnet.co.at) has joined #mythtv | |
| [06:48:53] | mrand (mrand!~mrand@ubuntu/member/mrand) has quit (Ping timeout: 248 seconds) | |
| [07:02:45] | dekarl1: | jya: is that something for you? http://code.mythtv.org/trac/attachment/ticket . . . x-typo.patch |
| [07:02:56] | dekarl1 is now known as dekarl | |
| [07:03:12] | jya: | houla… that's a blocker ! |
| [07:03:26] | jya: | weird.. I'm pretty sure I had fixed that one |
| [07:04:36] | dekarl: | I think most lines where it occurs contain "airplay", too so grep -i airplay will still work ;) |
| [07:07:46] | jya: | i know what happed, I had fixed it, and continued on some changes, nothing was working so I reverted |
| [07:12:33] | aloril (aloril!~aloril@dsl-tkubrasgw3-fe7ef900-153.dhcp.inet.fi) has quit (Ping timeout: 245 seconds) | |
| [07:19:27] | mrand (mrand!~mrand@ubuntu/member/mrand) has joined #mythtv | |
| [07:22:08] | jya: | don't tell me that I've spend the past 3 days writing a HLS client, when the new ffmpeg supports it !! |
| [07:25:55] | aloril (aloril!~aloril@84.249.126.153) has joined #mythtv | |
| [07:29:28] | dekarl: | ouch, maybe you can repurpose the code to a HLSRecorder? |
| [07:29:30] | jya: | Beirdo: VDA is all broken... |
| [07:30:38] | jya: | dekarl: I'm still not convinced that it fully supports it.. Someone told me it did… and as a consequence XBMC supported HLS playback.. just tried xbmc and it doesn't |
| [07:31:26] | jya: | just need to get myth to compile on a mac, and then I'll be able to try |
| [07:32:17] | dekarl: | jya: either way, looking forward to a nice HLS player, started to add some channels to my guide already :D |
| [07:33:48] | jya: | I believe I'm 90% done when it comes to parsing the m3u8 file, decoding the various segments find the segment to download and play, auto-adapt according to the network bandwidth… now just need to interface with the ringbuffer class |
| [07:37:07] | jya: | ByteIOContext doesn't exist anymore… need to find out what they replaced it with, seems to be AVIOContext |
| [07:37:11] | jya: | gotta love ffmpeg |
| [07:50:11] | rsiebert_ (rsiebert_!~quassel@g224248185.adsl.alicedsl.de) has joined #mythtv | |
| [07:53:20] | aloril (aloril!~aloril@84.249.126.153) has quit (Ping timeout: 272 seconds) | |
| [07:53:45] | rsiebert (rsiebert!~quassel@g224248096.adsl.alicedsl.de) has quit (Ping timeout: 260 seconds) | |
| [07:57:19] | cocoa117 (cocoa117!~cocoa117@188-222-31-239.zone13.bethere.co.uk) has quit (Read error: Operation timed out) | |
| [07:57:58] | cocoa117 (cocoa117!~cocoa117@188-222-31-239.zone13.bethere.co.uk) has joined #mythtv | |
| [08:06:16] | aloril (aloril!~aloril@84.249.126.153) has joined #mythtv | |
| [08:08:50] | Dr{Who} (Dr{Who}!~mathewss@c-98-232-143-236.hsd1.or.comcast.net) has quit (Ping timeout: 252 seconds) | |
| [08:18:15] | jya: | Beirdo: did you even try to compile raop ? :) |
| [08:27:56] | jya: | well, it plays crap :) |
| [08:28:30] | stuartm (stuartm!~stuartm@cpc1-derb9-0-0-cust441.8-3.cable.virginmedia.com) has joined #mythtv | |
| [08:28:31] | stuartm (stuartm!~stuartm@cpc1-derb9-0-0-cust441.8-3.cable.virginmedia.com) has quit (Changing host) | |
| [08:28:31] | stuartm (stuartm!~stuartm@mythtv/developer/stuartm) has joined #mythtv | |
| [08:44:01] | hashbang (hashbang!~alex@213-152-35-50.dsl.eclipse.net.uk) has joined #mythtv | |
| [09:00:22] | jya: | I'm pretty sure the change in avformatdecoder is not good either… |
| [09:07:11] | jya: | plus there's another issue at play, the audio decoded being passed on to the audio framework may not be 16 bytes aligned, this will cause a segfault if audio processing is on |
| [09:34:21] | jya: | looking at the ffmpeg avcodec_decode_audio3 ; they loop through each channel and via extended_data[x] to make a complete audio frame… I'm surprised that we get any audio anywhere the way it's been converted |
| [10:05:02] | Guest58244 (Guest58244!~mike@c-98-232-220-158.hsd1.or.comcast.net) has quit (Remote host closed the connection) | |
| [10:05:50] | mike (mike!~mike@c-98-232-220-158.hsd1.or.comcast.net) has joined #mythtv | |
| [10:06:16] | mike is now known as Guest67565 | |
| [10:14:46] | jya: | weird, the new ffmpeg is supposed to support HLS playback ; but the existing streamingbuffer obviously can't handle it properly... |
| [11:39:39] | markcerv_ (markcerv_!~mark@ns.shelfspace.com) has joined #mythtv | |
| [11:41:29] | gigem_ (gigem_!~david@pool-71-123-128-97.dllstx.fios.verizon.net) has joined #mythtv | |
| [11:41:29] | gigem_ (gigem_!~david@mythtv/developer/gigem) has joined #mythtv | |
| [11:41:29] | gigem_ (gigem_!~david@pool-71-123-128-97.dllstx.fios.verizon.net) has quit (Changing host) | |
| [11:41:35] | hashbang1 (hashbang1!~alex@213-152-35-50.dsl.eclipse.net.uk) has joined #mythtv | |
| [11:42:09] | kurre2 (kurre2!~tomimo@xdsl-83-150-88-111.nebulazone.fi) has quit (Ping timeout: 252 seconds) | |
| [11:42:09] | gigem (gigem!~david@mythtv/developer/gigem) has quit (Ping timeout: 252 seconds) | |
| [11:42:10] | Guest67565 (Guest67565!~mike@c-98-232-220-158.hsd1.or.comcast.net) has quit (Ping timeout: 252 seconds) | |
| [11:42:10] | markcerv (markcerv!~mark@ns.shelfspace.com) has quit (Ping timeout: 252 seconds) | |
| [11:42:10] | hashbang (hashbang!~alex@213-152-35-50.dsl.eclipse.net.uk) has quit (Ping timeout: 252 seconds) | |
| [11:42:22] | Guest67565 (Guest67565!~mike@c-98-232-220-158.hsd1.or.comcast.net) has joined #mythtv | |
| [11:45:32] | kurre2 (kurre2!~tomimo@xdsl-83-150-88-111.nebulazone.fi) has joined #mythtv | |
| [11:49:48] | knightr_ (knightr_!~knightr@69.165.170.178) has joined #mythtv | |
| [11:51:19] | jya_ (jya_!~jyavenard@120.148.99.165) has joined #mythtv | |
| [11:51:20] | jya_ (jya_!~jyavenard@120.148.99.165) has quit (Changing host) | |
| [11:51:20] | jya_ (jya_!~jyavenard@mythtv/developer/jya) has joined #mythtv | |
| [11:53:44] | knightr (knightr!~knightr@mythtv/developer/knightr) has quit (Ping timeout: 265 seconds) | |
| [11:55:11] | dekarl (dekarl!~dekarl@p4FE84F77.dip.t-dialin.net) has quit (Ping timeout: 265 seconds) | |
| [12:35:14] | dmfrey (dmfrey!~dmfrey@webdefence.cluster-h.websense.net) has joined #mythtv | |
| [13:07:42] | Goga777 (Goga777!~Goga777@2.95.137.147) has joined #mythtv | |
| [13:08:23] | Goga777 (Goga777!~Goga777@2.95.137.147) has quit (Read error: Connection reset by peer) | |
| [13:22:36] | jya_: | Beirdo: I've figured out why raop doesn't work. The AVFrame doesn't actually contains the decoded data, but a pointer to it. That pointer is the internal ffmpeg buffer, so next call to avcodec_decode_audio4 will overwrite it… so what you end up hearing is the last decoded audio frame… it must be copied first |
| [13:24:48] | jya_: | normally avcodec_decode_audio4 is designed to work with your own get_buffer routine.. Re-implementing the get-buffer would prevent having to make an extra copy |
| [13:30:06] | stichnot (stichnot!chatzilla@nat/intel/x-xfacvtbbobvtqgbl) has joined #mythtv | |
| [13:30:06] | stichnot (stichnot!chatzilla@nat/intel/x-xfacvtbbobvtqgbl) has quit (Changing host) | |
| [13:30:07] | stichnot (stichnot!chatzilla@mythtv/developer/stichnot) has joined #mythtv | |
| [13:38:53] | zombor_ (zombor_!~zombor_@kohana/developer/zombor) has joined #mythtv | |
| [14:05:20] | natanojl (natanojl!~jonatan@c83-252-237-63.bredband.comhem.se) has joined #mythtv | |
| [14:06:08] | zombor_ is now known as zombor | |
| [14:13:25] | Goga777 (Goga777!~Goga777@2.95.137.147) has joined #mythtv | |
| [14:23:30] | dmfrey (dmfrey!~dmfrey@webdefence.cluster-h.websense.net) has quit (Remote host closed the connection) | |
| [14:24:12] | dmfrey (dmfrey!~dmfrey@webdefence.cluster-h.websense.net) has joined #mythtv | |
| [14:30:19] | ben1066_ (ben1066_!~quassel@host86-147-240-100.range86-147.btcentralplus.com) has joined #mythtv | |
| [14:32:45] | ben1066 (ben1066!~quassel@host109-152-85-91.range109-152.btcentralplus.com) has quit (Ping timeout: 260 seconds) | |
| [14:39:25] | ben1066_ (ben1066_!~quassel@host86-147-240-100.range86-147.btcentralplus.com) has quit (Read error: Operation timed out) | |
| [14:41:01] | ben1066 (ben1066!~quassel@host86-151-190-177.range86-151.btcentralplus.com) has joined #mythtv | |
| [14:46:53] | Goga777 (Goga777!~Goga777@2.95.137.147) has quit (Remote host closed the connection) | |
| [14:47:08] | Goga777 (Goga777!~Goga777@2.95.137.147) has joined #mythtv | |
| [14:48:29] | Goga777 (Goga777!~Goga777@2.95.137.147) has quit (Client Quit) | |
| [14:49:05] | Goga777 (Goga777!~Goga777@2.95.137.147) has joined #mythtv | |
| [15:00:12] | solars (solars!~solars@clnet-kmu02-090.ikbnet.co.at) has quit (Read error: Operation timed out) | |
| [15:09:39] | stichnot (stichnot!chatzilla@mythtv/developer/stichnot) has quit (Ping timeout: 256 seconds) | |
| [15:11:28] | gregL (gregL!~greg@cpe-74-76-125-87.nycap.res.rr.com) has quit (Quit: Leaving) | |
| [15:17:50] | Goga777 (Goga777!~Goga777@2.95.137.147) has quit (Quit: Leaving) | |
| [15:28:06] | kurre2 (kurre2!~tomimo@xdsl-83-150-88-111.nebulazone.fi) has quit (Quit: leaving) | |
| [16:04:05] | joki (joki!~joki@p54861E87.dip.t-dialin.net) has quit (Ping timeout: 248 seconds) | |
| [16:04:10] | joki- (joki-!~joki@p54863008.dip.t-dialin.net) has joined #mythtv | |
| [16:04:26] | joki- is now known as joki | |
| [16:16:28] | jya_: | hum… looks like my HLS player isn't to be thrown away after all, ffmpeg doesn't handle selecting a stream according to the bandwidth available, or change stream on the fly as the bandwidth change… won't work well with live stream either. |
| [16:24:09] | stichnot (stichnot!chatzilla@nat/intel/x-eeuxqwomyxjexotv) has joined #mythtv | |
| [16:24:10] | stichnot (stichnot!chatzilla@nat/intel/x-eeuxqwomyxjexotv) has quit (Changing host) | |
| [16:24:10] | stichnot (stichnot!chatzilla@mythtv/developer/stichnot) has joined #mythtv | |
| [16:27:14] | ponyofdeath (ponyofdeath!~vladi@cpe-75-80-173-94.san.res.rr.com) has joined #mythtv | |
| [16:34:30] | kormoc (kormoc!~kormoc@mythtv/developer/kormoc) has joined #mythtv | |
| [16:51:16] | stichnot: | So this idea of adding thumbnails of surrounding frames to the cutlist editor is cool, but the performance is a bit frustrating. |
| [16:51:50] | stichnot: | I try to get it to use HW decoding for the thumbnails, and VDPAU appears to agree, but it's clear it's still being done in software. |
| [16:52:16] | stichnot: | which is a problem on an ION box |
| [16:53:24] | stichnot: | I generate the thumbnails one at a time in the Qt event loop, but responsiveness is still not good because each thumbnail may block the loop for 0.5–1.0s |
| [16:54:11] | stichnot: | so it's clear I need to set up a worker thread to do the heavy lifting |
| [16:54:48] | stichnot: | but still, it's annoying that the HW resources are taken up just for the main (paused) frame |
| [16:56:03] | stichnot: | so I'm considering tearing down the main player when the editor starts, and reconstructing it when the editor stops and playback resumes. |
| [16:56:24] | stichnot: | Are there any big problems with that approach? |
| [16:58:17] | jya_: | why would the pause frame take any resources? |
| [16:58:31] | jya_: | especially with vdpau, you can decode more than one frame at a time now |
| [17:00:06] | stichnot: | It seems I can't use the regular player to grab thumbnails, so I have to create a second MythPlayer instance. |
| [17:00:37] | jya_: | not with vdpau you can't grab no... |
| [17:01:10] | stichnot: | hmm, makes me wonder if things would be different with newer nvidia drivers |
| [17:01:50] | stichnot: | looks like I'm using 280.13, fwiw |
| [17:01:52] | jya_: | stichnot: not if you use vdpau decoding + rendering… rendering is done directly by the card. |
| [17:01:59] | NightMonkey (NightMonkey!~NightrMon@pdpc/supporter/professional/nightmonkey) has joined #mythtv | |
| [17:02:19] | jya_: | though, if i remember correctly, a little while back, when you entered to TV program ; playback in a thumbnail was disable at some stage |
| [17:02:38] | jya_: | markk could not get the playback to work there |
| [17:02:54] | stichnot: | yeah, it seems that PIP/PBP explicitly disables HW acceleration |
| [17:02:57] | jya_: | so he showed a single frame instead |
| [17:03:07] | jya_: | maybe it's worth checking what is done there |
| [17:03:21] | jya_: | and use the same method: there you would get the ability to capture the frame |
| [17:04:30] | stichnot: | good idea to check over there |
| [17:10:26] | stichnot: | I was also thinking that if the editor completely uses its own player infrastructure, I might actually be able to edit recordings on a remote X server over ssh, using the beefy backend, etc. |
| [17:17:56] | jya_: | i wouldn't have thought that editing a recording would take much time |
| [17:21:06] | stichnot: | it depends on how good commflagging was |
| [17:23:53] | stichnot: | with almost-perfect commflagging, I can do a 1-hour recording in 30–60 seconds (U.S. broadcast networks, 5–6 commercial breaks, edited length 41–42 minutes) |
| [17:24:21] | stichnot: | it can take 2 minutes if commflagging did a mediocre job |
| [17:25:38] | stichnot: | this is doing it as a "power user", and it becomes incredibly tedious if the delay in jumping to a new frame gets too long |
| [17:27:02] | stichnot: | editing an hdpvr recording on an ION is an example — it can take 1 second to jump elsewhere in the recording, since decoding seems to be up to 120fps and keyframes are 128 frames apart |
| [17:32:25] | jya_: | i see |
| [17:32:54] | stichnot: | btw, I use the editor a lot because I prefer to chop the ads before starting to watch |
| [17:33:13] | stichnot: | my wife complains that now we don't know what movies are playing :) |
| [17:37:41] | jya_: | i never do |
| [17:37:56] | jya_: | commflag is so unreliable here, so i just keep 4 minutes |
| [17:38:02] | jya_: | does the job |
| [17:41:18] | solars (solars!~solars@194.208.132.118) has joined #mythtv | |
| [17:41:35] | stichnot: | yeah, without reasonable commflag as a starting point, I think I would rarely bother to edit, except maybe to trim the start and end |
| [17:46:29] | stichnot: | "keep 4 minutes" — did you mean jump 4 minutes? |
| [17:48:22] | jya_: | yep |
| [17:49:28] | stichnot: | if only commercial break lengths were consistent here... |
| [17:51:21] | danielk22: | jya: In ffmpeg, is "LPCM" just uncompressed audio? i.e. 16,24,32 bit linear samples at whatever bitrate? |
| [17:52:07] | jya_: | danielk22: yes |
| [17:52:44] | jya_: | why do you ask? |
| [17:53:29] | jya_: | danielk22: do you have much experience with ffmpeg? |
| [17:53:48] | danielk22: | I want to write out some audio of unknown codec to disk for analysis.. right now I'm writing out a raw pcm file, but I don't get an PTS with that. |
| [17:54:50] | danielk22: | jya_: I've submitted some patches, but I try to stay away from the internals as much as possible. |
| [17:54:51] | jya_: | danielk22: you could quite easily write an AudioOutput subclass that write everything it receives to disk |
| [17:55:37] | jya_: | you specify the audio format you support (S16) and you get raw pcm on disk |
| [17:55:43] | danielk22: | jya_: Ideally I can just wrap it up in mpeg-ts as a private stream. |
| [17:55:57] | jya_: | i did one once… don't know where that code is |
| [17:56:05] | danielk22: | jya_: It's 24 bit so I'll probably use S32. |
| [17:56:17] | jya_: | danielk22: that too :) |
| [17:57:17] | jya_: | http://pastebin.com/M5dXf03A |
| [17:58:00] | jya_: | i added a quick hack to the streambuffer class, so if it sees that it ends with .m3u8, it changes the protocol from http to hls+http |
| [17:58:14] | jya_: | which will automatcally read a http live stream |
| [17:58:58] | jya_: | but it's pretty poor feature wise... |
| [17:59:09] | jya_: | so i wonder what is required to use a "demuxer" |
| [18:06:46] | amejia_ (amejia_!~andres@129.174.97.130) has joined #mythtv | |
| [18:06:46] | amejia_ (amejia_!~andres@xbmc/staff/amejia) has joined #mythtv | |
| [18:06:46] | amejia_ (amejia_!~andres@129.174.97.130) has quit (Changing host) | |
| [18:17:15] | cattelan (cattelan!~cattelan@c-66-41-26-220.hsd1.mn.comcast.net) has quit (Quit: Terminated with extreme prejudice - dircproxy 1.2.0) | |
| [18:31:55] | ben1066 (ben1066!~quassel@host86-151-190-177.range86-151.btcentralplus.com) has quit (Ping timeout: 260 seconds) | |
| [18:37:06] | pheld (pheld!~heldal@cl-5.osl-01.no.sixxs.net) has joined #mythtv | |
| [18:41:18] | ben1066 (ben1066!~quassel@host86-160-30-34.range86-160.btcentralplus.com) has joined #mythtv | |
| [18:41:45] | cattelan (cattelan!~cattelan@c-66-41-26-220.hsd1.mn.comcast.net) has joined #mythtv | |
| [18:41:49] | Beirdo: | kiss goodbye to mythffplay by default. |
| [18:42:36] | Beirdo: | we have an issue... it enables sdl by default, but on Winblows, we have SDL, but we don't *want* it for our programs, the only thing that requires it is ffplay |
| [18:42:48] | Beirdo: | so, if you want ffplay, add --enable-sdl :) |
| [18:43:48] | stuartm: | iirc mythmusic uses SDL for visualisers |
| [18:44:14] | Beirdo: | I beleive that was ripped out in 0.25 |
| [18:44:47] | Beirdo: | http://www.mythtv.org/wiki/Release_Notes_-_0. . . . site_Changes |
| [18:45:32] | Beirdo: | if that's incorrect, we can always revert and pound on it more ;) |
| [18:45:47] | kurre2 (kurre2!~tomimo@xdsl-83-150-88-111.nebulazone.fi) has joined #mythtv | |
| [18:46:43] | stuartm: | looks like synaesthesia is still using it |
| [18:47:02] | stuartm: | it's not using it through ffmpeg, but it is using it |
| [18:47:17] | Beirdo: | hmm. |
| [18:47:57] | Beirdo: | Well, we'll see what the buildbot thinks of it :) It's not a problem to take another whack at this if it misbehaves. |
| [18:48:23] | MythBuild: | build #1094 of master-vista-mingw-32bit is complete: Exception [6exception git] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . /builds/1094 blamelist: Gavin Hurlbut <ghurlbut@mythtv.org > |
| [18:48:29] | stuartm: | hmm, maybe that code is commented out though, grep isn't showing enough context |
| [18:48:38] | Beirdo: | OK, THAT doesn't count |
| [18:48:46] | Beirdo: | I forgot to cd out of the working dir |
| [18:49:14] | Beirdo: | MythBuild: force build master-vista-mingw-32bit |
| [18:49:14] | MythBuild: | build forced [ETA 32m29s] |
| [18:49:14] | MythBuild: | I'll give a shout when the build finishes |
| [18:49:28] | stuartm: | ah, yeah, thought that might be it – the SDL using code is still there but commented out |
| [18:49:38] | stuartm: | so it's not actually using SDL anymore |
| [18:51:36] | Beirdo: | cool :) |
| [18:53:04] | stuartm: | I think that's the second time I've done that, spotted the SDL usage via grep only to realise afterwards that it's in a "if 0" block |
| [18:55:52] | stuartm: | does anyone know how http://svn.mythtv.org/trac/ticket/10657 is our problem? |
| [18:56:24] | stuartm: | we don't supply the init scripts or configs |
| [18:59:13] | danielk22: | stuartm: If we don't respond to the signal properly it is our problem. |
| [19:03:06] | tomimo (tomimo!~kurre@xdsl-83-150-88-111.nebulazone.fi) has joined #mythtv | |
| [19:03:51] | Beirdo: | we should remove that commented code to keep it from recurring :) |
| [19:04:22] | stuartm: | agreed, but are we sure that's the case here? The backend responds correctly here to the usual signals |
| [19:04:24] | Beirdo: | danielk22: I'd hold off on doing anything on that ticket |
| [19:04:43] | Beirdo: | I have a redo of the logging ready to go, including proper signal handlers |
| [19:04:44] | danielk22: | It looks like ProgramInfoUpdater::run() had just sent an update and was waiting to see if there were more updates pending before shutting down. |
| [19:05:22] | Beirdo: | turns out (according to Qt documents) we were very naughty |
| [19:05:35] | danielk22: | ? |
| [19:05:46] | Beirdo: | http://doc.qt.nokia.com/4.7-snapshot/unix-signals.html |
| [19:06:08] | Beirdo: | You can't call Qt functions from Unix signal handlers. |
| [19:06:14] | Beirdo: | to quote the very first line |
| [19:06:32] | Beirdo: | and that's exactly what we were doing... not sure how that could affect things, but... |
| [19:06:42] | stuartm: | well, maybe you're not supposed to, but you can do it ;) |
| [19:06:49] | Beirdo: | so I have redone our signal handlers to closely match theirs |
| [19:07:04] | Beirdo: | as crappy as it is, and they do work |
| [19:07:18] | Beirdo: | including one on SEGV that will delay, then re-raise SEGV |
| [19:08:21] | Beirdo: | so, once that's merged, if the problem persists, then we can look at making sure the threads are behaving. Of course, you can do it in the opposite order if you wish :) |
| [19:08:41] | Beirdo: | what signal are they sending from the init script, do we know? |
| [19:09:35] | danielk22: | Beirdo: I dunno, but I use --retry TERM/30/ABRT/5/KILL/5 |
| [19:09:55] | Beirdo: | hehe, KILL will definitely stop it |
| [19:10:26] | Beirdo: | TERM and INT have handlers (in the new and old code) |
| [19:10:38] | Beirdo: | ABRT, I'm treating the same as SEGV... delay and exit |
| [19:11:05] | Beirdo: | that way the logging should flush (unless the code died in the logging thread, in which case we are hosed) |
| [19:11:21] | danielk22: | Well, TERM is supposed to take it down, the abort is to try to get a core if we don't respond to SIGTERM, then KILL in case all else fails. |
| [19:11:29] | Beirdo: | right |
| [19:11:34] | Beirdo: | makes sense |
| [19:11:51] | Beirdo: | I woulda probably done TERM->SEGV->KILL for the same reason |
| [19:12:01] | Beirdo: | it's all the same |
| [19:12:59] | Beirdo: | anyways, once we have the ffmpeg sync stuff compiling on all slaves again, I was looking at merging the logging changes down too, but only want to deal with one pile at a time :) |
| [19:13:21] | danielk22: | Beirdo: the no Qt stuff in signal handlers makes sense. But I'm fairly certain that isn't the cause of any of our shutdown problems. |
| [19:13:28] | Beirdo: | we are down to PPC, OSX, Windows |
| [19:13:33] | Beirdo: | agreed |
| [19:13:43] | Beirdo: | but who knows :) |
| [19:14:49] | Beirdo: | it's up to you, of course. Just would be a shame to waste time on it if it's fixed by something that's ready to go. If not, well... a couple days shouldn't kill it, right? :) |
| [19:18:10] | MythBuild: | Hey! build master-vista-mingw-32bit #1095 is complete: Failure [4failed compile] |
| [19:18:10] | ** MythLogBot http://code.mythtv.org/trac/ticket/1095 ** | |
| [19:18:10] | MythBuild: | Build details are at http://code.mythtv.org/buildbot/builders/mast . . . /builds/1095 |
| [19:18:51] | Beirdo: | come ON |
| [19:19:01] | Beirdo: | I told it to disable SDL, dammit |
| [19:24:02] | stuarta: | hahaha |
| [19:24:55] | Beirdo: | I will choke the living daylights out of this configure script, I tell ya |
| [19:25:21] | Beirdo: | I have an idea though |
| [19:25:44] | Beirdo: | first, a fix for the ppc build |
| [19:32:32] | MythBuild: | build #1096 of master-vista-mingw-32bit is complete: Exception [6exception git] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . /builds/1096 blamelist: Gavin Hurlbut <ghurlbut@mythtv.org > |
| [19:32:48] | Beirdo: | I hate windows |
| [19:32:59] | Beirdo: | MythBuild: force build master-vista-mingw-32bit |
| [19:32:59] | MythBuild: | build forced [ETA 32m29s] |
| [19:33:00] | MythBuild: | I'll give a shout when the build finishes |
| [19:33:23] | Beirdo: | if you are sitting in the build dir in a terminal, it can't copy over it |
| [19:34:55] | MythBuild: | build #2527 of master-freebsd-64bit is complete: Failure [4failed compile core] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . /builds/2527 blamelist: Gavin Hurlbut <ghurlbut@mythtv.org > |
| [19:35:38] | Beirdo: | grrr. |
| [19:36:24] | stuarta: | what happened to the commits going into #mythtv-commits |
| [19:36:29] | stuarta: | from the bot |
| [19:36:42] | Beirdo: | the IRC bot we have doesn't work |
| [19:36:55] | Beirdo: | I can make it go there from github if we want in the meantime |
| [19:37:14] | stuarta: | does it just need some love to work on alcor? |
| [19:37:33] | Beirdo: | it needs a new bot. There is no standard one |
| [19:37:41] | Beirdo: | it's on my list :) |
| [19:37:46] | stuarta: | :) |
| [19:40:01] | sraue (sraue!~stephan@xbmc/staff/sraue) has quit (Quit: Ex-Chat) | |
| [19:41:47] | stuartm: | Beirdo: re-enabling the github one will suffice for now and the odd missed commit doesn't matter so much for the irc channel |
| [19:41:56] | stuartm: | Beirdo: I'll do it |
| [19:41:59] | Beirdo: | yeah, true |
| [19:42:01] | Beirdo: | K :) |
| [19:42:43] | Beirdo: | I don't know why that didn't cross my mind last time it was mentioned. Too many things bouncing around in my head, I guess |
| [19:42:50] | sraue (sraue!~stephan@xbmc/staff/sraue) has joined #mythtv | |
| [19:51:19] | MythBuild: | build #3747 of master-linux-64bit is complete: Failure [4failed compile core] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . /builds/3747 blamelist: Gavin Hurlbut <ghurlbut@mythtv.org > |
| [19:56:17] | MythBuild: | Hey! build master-vista-mingw-32bit #1097 is complete: Failure [4failed compile] |
| [19:56:17] | ** MythLogBot http://code.mythtv.org/trac/ticket/1097 ** | |
| [19:56:17] | MythBuild: | Build details are at http://code.mythtv.org/buildbot/builders/mast . . . /builds/1097 |
| [20:00:41] | Beirdo: | more missing crap in configure. :) It is decidedly not my day today |
| [20:01:02] | stuarta: | do i need to throw something sharp at the osx builder? |
| [20:01:26] | MythBuild: | build #2225 of master-debian-stable-64bit is complete: Failure [4failed compile core] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . /builds/2225 blamelist: Gavin Hurlbut <ghurlbut@mythtv.org > |
| [20:01:40] | stuarta: | 1 step forwards, 2 back |
| [20:01:52] | Beirdo: | yup |
| [20:02:18] | Beirdo: | this just makes me all the more determined to separate our configure from ffmpeg's |
| [20:02:52] | MythBuild: | build #3477 of master-linux-32bit is complete: Failure [4failed compile core] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . /builds/3477 blamelist: Gavin Hurlbut <ghurlbut@mythtv.org > |
| [20:03:04] | Beirdo: | not sure where teh OSX was blocked... pthreads or something? |
| [20:03:11] | stichnot (stichnot!chatzilla@mythtv/developer/stichnot) has quit (Ping timeout: 256 seconds) | |
| [20:03:12] | stuarta: | nah, that was debian |
| [20:03:21] | Beirdo: | oooh, right, needing -O3 |
| [20:03:30] | stuarta: | its the odd register thing, yeah -O3 |
| [20:03:49] | stichnot (stichnot!~chatzilla@mythtv/developer/stichnot) has joined #mythtv | |
| [20:04:13] | Beirdo: | hmm, the simplest thing to do I guess is check the config.mak |
| [20:04:44] | Beirdo: | and if you want to see the compile line, you can remove CC from the "BRIEF" definition in common.mak in FFmpeg |
| [20:05:32] | Beirdo: | I'm test compiling AGAIN with the configure change. |
| [20:06:19] | Beirdo: | and putting on loud music to drown out my cow-workers |
| [20:07:04] | amejia_ (amejia_!~andres@xbmc/staff/amejia) has joined #mythtv | |
| [20:08:33] | Beirdo: | OK, that should at least fix the 64bit linux (as I just compiled it there) |
| [20:12:50] | stuarta: | that's one sea of red on the buildbot waterfall output |
| [20:13:07] | Beirdo: | yeah. :( |
| [20:13:48] | Beirdo: | grrrr |
| [20:15:15] | stuarta: | osx config.mak sent |
| [20:16:02] | MythBuild: | build #3748 of master-linux-64bit is complete: Success [3build successful] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . /builds/3748 |
| [20:16:49] | gregL (gregL!~greg@cpe-74-76-125-87.nycap.res.rr.com) has joined #mythtv | |
| [20:16:59] | Beirdo: | yay... getting better |
| [20:17:08] | Beirdo: | wagnerrp: you in? |
| [20:17:38] | Beirdo: | oh no. |
| [20:17:45] | Beirdo: | Obama is in Seattle tomorrow. |
| [20:18:08] | Beirdo: | right next door to work... I think I should work from home, and avoid the idiocy |
| [20:20:50] | stichnot: | no political discussions please |
| [20:20:52] | stichnot: | :) |
| [20:20:56] | MythBuild: | build #2226 of master-debian-stable-64bit is complete: Success [3build successful] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . /builds/2226 |
| [20:21:06] | Beirdo: | this isn't politics, it's transit bugger-ups |
| [20:21:37] | Beirdo: | if the president is in town, the busses get messed up, and the protesters are out in force |
| [20:21:43] | Beirdo: | doesn't matter what president |
| [20:21:44] | MythBuild: | build #3478 of master-linux-32bit is complete: Success [3build successful] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . /builds/3478 |
| [20:22:01] | stichnot: | just joking about the "idiocy" comment |
| [20:22:07] | Beirdo: | hehe :) |
| [20:22:20] | Beirdo: | trust me, in Seattle... Obama is generally well-liked |
| [20:22:31] | Beirdo: | not like Bush :) |
| [20:22:37] | Beirdo: | anyways, moving on |
| [20:28:15] | Beirdo: | no wagnerrp? |
| [20:29:15] | Beirdo: | so now I know when he sleeps :) |
| [20:40:22] | wahrhaft (wahrhaft!~quassel@cpe-24-210-71-26.columbus.res.rr.com) has joined #mythtv | |
| [20:41:34] | amejia__ (amejia__!~andres@xbmc/staff/amejia) has joined #mythtv | |
| [20:42:24] | Beirdo: | this is just rediculous |
| [20:42:26] | ghoti_ (ghoti_!~paul@scratch.it.ca) has joined #mythtv | |
| [20:42:28] | brtb (brtb!~brtb@linode.brtb.net) has joined #mythtv | |
| [20:42:39] | Beirdo: | all this for a bloody byte-swap?! |
| [20:44:12] | MythBuild: | build #2531 of master-freebsd-64bit is complete: Success [3build successful] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . /builds/2531 |
| [20:46:58] | amejia_ (amejia_!~andres@xbmc/staff/amejia) has quit (Remote host closed the connection) | |
| [20:46:58] | amejia (amejia!~andres@xbmc/staff/amejia) has quit (Remote host closed the connection) | |
| [20:47:01] | brtb_ (brtb_!~brtb@linode.brtb.net) has quit (Write error: Broken pipe) | |
| [20:47:01] | kurre2 (kurre2!~tomimo@xdsl-83-150-88-111.nebulazone.fi) has quit (Write error: Broken pipe) | |
| [20:47:02] | ghoti (ghoti!~paul@scratch.it.ca) has quit (Write error: Broken pipe) | |
| [20:47:04] | sutula (sutula!sutula@nat/hp/x-sqaesklczhucdlpx) has quit (Excess Flood) | |
| [20:47:05] | wahrhaft_ (wahrhaft_!~quassel@cpe-24-210-71-26.columbus.res.rr.com) has quit (Remote host closed the connection) | |
| [20:47:09] | Beirdo: | yay |
| [20:47:19] | Beirdo: | now I just need to get Winblows happy with it. |
| [20:47:24] | sutula (sutula!sutula@nat/hp/x-rxmgbatzxkgbynte) has joined #mythtv | |
| [20:47:33] | amejia_ (amejia_!~andres@129.174.97.130) has joined #mythtv | |
| [20:47:33] | amejia_ (amejia_!~andres@xbmc/staff/amejia) has joined #mythtv | |
| [20:47:33] | amejia_ (amejia_!~andres@129.174.97.130) has quit (Changing host) | |
| [20:47:51] | dmfrey (dmfrey!~dmfrey@webdefence.cluster-h.websense.net) has quit (Quit: Ex-Chat) | |
| [20:48:47] | kurre2 (kurre2!~tomimo@xdsl-83-150-88-111.nebulazone.fi) has joined #mythtv | |
| [20:54:26] | MythBuild: | build #2453 of master-linux-ppc is complete: Success [3build successful] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . /builds/2453 |
| [20:55:43] | Beirdo: | YAY |
| [20:56:54] | ** stuartm waits for all the other builds to start failing instead ** | |
| [20:59:29] | stichnot_ (stichnot_!chatzilla@mythtv/developer/stichnot) has joined #mythtv | |
| [21:00:13] | GreyFoxx (GreyFoxx!~greg@out.of.phaze.org) has quit (Ping timeout: 260 seconds) | |
| [21:00:34] | GreyFoxx (GreyFoxx!~greg@out.of.phaze.org) has joined #mythtv | |
| [21:00:53] | stichnot (stichnot!~chatzilla@mythtv/developer/stichnot) has quit (Ping timeout: 245 seconds) | |
| [21:00:55] | stichnot_ is now known as stichnot | |
| [21:03:12] | _charly_ (_charly_!kroseneg@sunrise.schmidham.net) has quit (Ping timeout: 260 seconds) | |
| [21:03:17] | stuartm: | heh, the coverity defect browser is awful, by comparison my cppcheck report is a work of art |
| [21:05:00] | stichnot: | anything interesting so far? |
| [21:09:50] | zombor_ (zombor_!~zombor_@kohana/developer/zombor) has joined #mythtv | |
| [21:10:20] | cattelan is now known as cattelan_away | |
| [21:11:27] | ben1066 (ben1066!~quassel@host86-160-30-34.range86-160.btcentralplus.com) has quit (Read error: Connection reset by peer) | |
| [21:11:27] | zombor (zombor!~zombor_@kohana/developer/zombor) has quit (Ping timeout: 260 seconds) | |
| [21:11:27] | ben1066 (ben1066!~quassel@86.160.30.34) has joined #mythtv | |
| [21:11:27] | foxbuntu (foxbuntu!~foxbuntu@ubuntu/member/foxbuntu) has quit (Ping timeout: 260 seconds) | |
| [21:12:01] | cattelan_away is now known as cattelan | |
| [21:14:30] | foxbuntu (foxbuntu!~foxbuntu@174-30-104-225.desm.qwest.net) has joined #mythtv | |
| [21:14:37] | foxbuntu (foxbuntu!~foxbuntu@174-30-104-225.desm.qwest.net) has quit (Changing host) | |
| [21:14:38] | foxbuntu (foxbuntu!~foxbuntu@ubuntu/member/foxbuntu) has joined #mythtv | |
| [21:16:53] | stuartm: | hard to tell, it's that difficult to read – there's no severity assigned to bugs by default, so you've got 900+ defects in no particular order, most of which are trivial warnings rather than real issues |
| [21:18:32] | dissent (dissent!~Ponzo@ponzo.ponzo.net) has joined #mythtv | |
| [21:18:34] | highzeth (highzeth!~hz@hoiseth.no) has quit (Ping timeout: 260 seconds) | |
| [21:18:40] | highzeth (highzeth!~hz@hoiseth.no) has joined #mythtv | |
| [21:19:34] | stuartm: | there are filters, I'm just figuring them out now |
| [21:21:48] | stuartm: | it's checking external code so a decent number of those 'issues' are in someone elses code, e.g. ffmpeg |
| [21:24:54] | sraue (sraue!~stephan@xbmc/staff/sraue) has quit (Remote host closed the connection) | |
| [21:25:37] | vallor (vallor!~Ponzo@pdpc/supporter/monthlygold/vallor) has quit (Ping timeout: 260 seconds) | |
| [21:25:38] | rsiebert (rsiebert!~quassel@g224248185.adsl.alicedsl.de) has joined #mythtv | |
| [21:25:38] | rsiebert_ (rsiebert_!~quassel@g224248185.adsl.alicedsl.de) has quit (Ping timeout: 260 seconds) | |
| [21:25:38] | J-e-f-f-A (J-e-f-f-A!~J-e-f-f-A@unaffiliated/j-e-f-f-a) has quit (Ping timeout: 260 seconds) | |
| [21:25:43] | J-e-f-f-A (J-e-f-f-A!~J-e-f-f-A@173.48.117.185) has joined #mythtv | |
| [21:26:33] | stuartm: | Beirdo: system-unix.cpp:963 – Variable "fd" is not closed or saved in function "dup2". |
| [21:27:57] | stuartm: | which is a slightly mis-leading synopsis, basically it's saying the file descriptor isn't closed under some circumstances |
| [21:27:59] | sraue (sraue!~stephan@xbmc/staff/sraue) has joined #mythtv | |
| [21:28:05] | Beirdo: | I don't think we particularly care in that situation |
| [21:28:19] | J-e-f-f-A (J-e-f-f-A!~J-e-f-f-A@173.48.117.185) has quit (Changing host) | |
| [21:28:19] | J-e-f-f-A (J-e-f-f-A!~J-e-f-f-A@unaffiliated/j-e-f-f-a) has joined #mythtv | |
| [21:28:29] | Beirdo: | as it's not supposed to be closed |
| [21:28:46] | Beirdo: | it's supposed to stay open, and take over stderr (hence the dup2) |
| [21:30:52] | stuartm: | two possible resolutions then, 'false positive' or 'intentional' |
| [21:31:01] | Beirdo: | intentional, I think |
| [21:31:11] | Beirdo: | ah hell. fire drill?! |
| [21:31:24] | Beirdo: | didn't I get enough of these in elementary school, FFS? |
| [21:31:32] | Beirdo: | so much for getting work done. |
| [21:35:29] | Steve_Goodey (Steve_Goodey!~steve@host86-182-4-91.range86-182.btcentralplus.com) has quit (Remote host closed the connection) | |
| [21:43:16] | MythBuild: | build #1101 of master-vista-mingw-32bit is complete: Success [3build successful] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . /builds/1101 |
| [21:43:24] | Beirdo: | YAY! |
| [21:43:34] | Beirdo: | now for that pesky OSX :) |
| [21:43:43] | danielk22: | Beirdo: I used to work somewhere with a daily 15 minute fire drill. My head still hurts just thinking about it. |
| [21:43:51] | Beirdo: | daily?! |
| [21:43:53] | Beirdo: | jeez |
| [21:44:16] | Beirdo: | that's way excessive unless you have Arsonists Anonymous meetings in the building |
| [21:47:13] | jya_ (jya_!~jyavenard@mythtv/developer/jya) has quit (Quit: jya_) | |
| [21:48:31] | gregL (gregL!~greg@cpe-74-76-125-87.nycap.res.rr.com) has quit (Quit: Leaving) | |
| [21:49:00] | andreax (andreax!~andreaz@p5089F59C.dip.t-dialin.net) has joined #mythtv | |
| [21:49:14] | sunkan (sunkan!~sunkan@alva.zappa.cx) has quit (Quit: leaving) | |
| [21:50:02] | stichnot: | If I want a helper thread to do work in the background, is the following correct? Define a subclass of MThread; override the run() method; create an instance of my subclass; a new thread is automatically created which executes the run() method; both threads point to the same instance of my subclass. |
| [21:53:37] | gigem_: | stuartm: Is there an existing MythUIScreenType (or parent) signal to which I can connect to get notified when it closes or do I need to add my own new one? As best I can tell, Exiting() is the best available choice, but that's run from the destructor and not from Close(). |
| [21:53:53] | gigem_ is now known as gigem | |
| [21:56:35] | zombor_ (zombor_!~zombor_@kohana/developer/zombor) has quit (Remote host closed the connection) | |
| [21:56:58] | stuartm: | gigem: MythScreenType::Exiting() |
| [21:57:20] | stuartm: | gigem: sorry, didn't read the complete question |
| [21:59:38] | stuartm: | Exiting is the only one currently available, it will be called ultimately when Close() is used, provided we've not requested that the screen be kept instead of deleted when it's removed from the stack |
| [22:01:01] | gigem: | Okany, that's sounds like it will work. I'll give it a try and come back if it doesn't. |
| [22:01:38] | stuartm: | there would be a small delay between Close() and the screen actually being deleted, but it should be very short |
| [22:02:47] | stuartm: | we're using Exiting() in a couple of places with no complaints, but it all depends on what you're trying to do :) |
| [22:03:42] | gigem: | Yeah, that "very short" is what I was hoping isn't "not short enough". We'll see. It will probably be okay. |
| [22:08:20] | solars (solars!~solars@194.208.132.118) has quit (Read error: Operation timed out) | |
| [22:09:10] | stuartm: | it's the time between Close() and the screen actually disappearing, that's <1 second, some re-implementations of MythScreenType send their own signals either to pass results back to a parent screen or to trigger outside events, that might be more suitable in your case? |
| [22:10:40] | stuartm: | we can always add a new signal sent from Close(), but not all instances of MythScreenType() are closed using Close() but instead by calling PopScreen() directly on the stack |
| [22:11:47] | gregL (gregL!~greg@cpe-74-76-125-87.nycap.res.rr.com) has joined #mythtv | |
| [22:14:47] | sraue (sraue!~stephan@xbmc/staff/sraue) has quit (Remote host closed the connection) | |
| [22:16:34] | sraue (sraue!~stephan@xbmc/staff/sraue) has joined #mythtv | |
| [22:17:21] | gigem: | stuartm: The more I think about it, a new signal might be better. While Exiting might work for my immediate needs, I don't think it will work for a more complicated case I will also want to handle. This is more of a Qt question, but is there any guarantee that slots are called in the same order as the signals? |
| [22:18:20] | stuartm: | yes, there are called immediately, basically just a more sophisticated type of callback |
| [22:20:20] | stuartm: | well you _can_ use queued signals, but they are still handled in order |
| [22:21:46] | gigem: | Good. I figured that was the case, but wasn't sure. |
| [22:22:04] | stuartm: | and you have to specify that you want the signal to be queued when calling connect(), the default is an 'direct' signal which interrupts the flow until the signal is sent |
| [22:24:22] | stuarta: | haha. victory over libvirt |
| [22:26:53] | stuarta: | once this build finishes i can check i have native ipv6 to my kvm vm |
| [22:27:11] | Beirdo: | oh? |
| [22:27:52] | stuarta: | well, it's all about defining the network correctly |
| [22:28:01] | stuarta: | and in theory it should work |
| [22:28:47] | stuarta: | however dnsmasq (which is used to do the dns & dhcp for the virbr interface) fails to start with the cryptic exit status 2 |
| [22:29:30] | stuarta: | in ipv4 this was normally due something else binding to the dns port globally |
| [22:29:55] | jcarlos (jcarlos!~quassel@85.137.99.76.dyn.user.ono.com) has quit (Read error: Connection reset by peer) | |
| [22:30:00] | stuarta: | in ipv6 the static ip addr i gave to the interface was marked as tentative, and you cannot bind to an address that is marked as tentative |
| [22:30:11] | stuarta: | thus it failed |
| [22:30:27] | stuarta: | a few sysctl tweaks later, and it works again |
| [22:30:51] | stuarta: | then there's a build issue with libvirt on debian/ubuntu that means it looks for radvd in / |
| [22:31:07] | stuarta: | quick symlink fixed that, and it's all good |
| [22:31:07] | jcarlos (jcarlos!~quassel@85.137.99.76.dyn.user.ono.com) has joined #mythtv | |
| [22:31:58] | Beirdo: | hehe |
| [22:32:49] | stuarta: | that's been bugging me since the weekend |
| [22:33:44] | Beirdo: | so we are down to just the OSX build failing |
| [22:33:56] | stuarta: | \o/ |
| [22:33:59] | Beirdo: | and the cflags in that config.mak indicate -O3 |
| [22:34:49] | ** stuarta does the happy dance ** | |
| [22:35:19] | Beirdo: | so I'm not quite sure why OSX is being so... Apple. |
| [22:37:46] | Beirdo: | I woulda thunk we could find recent reports of this, but no luck so far |
| [22:44:26] | Beirdo: | https://gist.github.com/1528236 |
| [22:44:32] | Beirdo: | that looks pretty similar |
| [22:44:54] | obo (obo!~obo@rockbox/developer/obo) has quit (Ping timeout: 260 seconds) | |
| [22:45:37] | obo (obo!~obo@70-140-190-120.lightspeed.hstntx.sbcglobal.net) has joined #mythtv | |
| [22:45:38] | obo (obo!~obo@70-140-190-120.lightspeed.hstntx.sbcglobal.net) has quit (Changing host) | |
| [22:45:38] | obo (obo!~obo@rockbox/developer/obo) has joined #mythtv | |
| [22:46:00] | stuarta: | yup that certainly looks the same |
| [22:46:45] | Beirdo: | seem to be no recent changes to the source file |
| [22:46:54] | Beirdo: | most recent: 2011-11–09 |
| [22:47:06] | Beirdo: | which is well before 2012-03–31 |
| [22:48:04] | Beirdo: | I see a suggestion to --disable-mmx |
| [22:51:20] | stuartm: | stichnot: coverity is suggesting that SubtitleFormat::CreateProviderDefault() is leaking 'font' and there do seem to be some cases where that happens |
| [22:53:45] | stichnot: | ok, I'll check |
| [22:54:15] | Dr{Who} (Dr{Who}!~mathewss@npv.nutech.com) has joined #mythtv | |
| [22:56:41] | stuartm: | danielk22: it also seems to think that we should be deleting the object return by hdhomerun_device_create_from_str() which is used in a few places in CardUtil |
| [22:59:15] | stuarta: | right, bed time |
| [23:05:53] | stichnot: | stuartm: very clever. looks like SubtitleFormat::Load() should be deleting negFont before it returns |
| [23:08:38] | hashbang1 (hashbang1!~alex@213-152-35-50.dsl.eclipse.net.uk) has quit (Quit: Leaving.) | |
| [23:13:04] | jya_ (jya_!~jyavenard@120.148.99.165) has joined #mythtv | |
| [23:13:20] | jya_ (jya_!~jyavenard@120.148.99.165) has quit (Changing host) | |
| [23:13:20] | jya_ (jya_!~jyavenard@mythtv/developer/jya) has joined #mythtv | |
| [23:15:45] | jya_ (jya_!~jyavenard@mythtv/developer/jya) has quit (Client Quit) | |
| [23:18:48] | Captain_Murdoch (Captain_Murdoch!~Captain_M@mythtv/developer/CaptainMurdoch) has quit (Ping timeout: 245 seconds) | |
| [23:21:47] | jya: | morning all |
| [23:22:20] | jya: | Beirdo: the mac version compile here… I wouldn't bother trying to make the OS X buildbot compile if there's an issue.. it needs to be migrated to the new build script |
| [23:22:41] | jya: | I haven't tried compiling myth without compile-type=debug |
| [23:22:58] | Beirdo: | hmmm, there is that. |
| [23:23:23] | jya: | in the previous case, without -g , llvm segfault on a ffmpeg function (in libswscale) |
| [23:23:51] | Beirdo: | yeah, this is dying on an MMX function in libswscale |
| [23:23:59] | Beirdo: | lack of registers |
| [23:24:19] | Beirdo: | we could try with --disable-mmx for now |
| [23:24:24] | jya: | the mac build is rather fragile, so I much prefer that it compiles for me and the other group building OSX, than fix an old builder that isn't used anymore by any packager and wth the potential on breaking something else |
| [23:24:37] | jya: | I can't compile with -O3 on the mac |
| [23:24:46] | Beirdo: | fair enough |
| [23:25:06] | jya: | llvm optimise too much and remove some code, so it fails at linking some of the filters |
| [23:25:15] | Beirdo: | well, I'll leave it to you and stuarta to get the OSX slave to a happy place, I guess |
| [23:25:31] | jya: | Beirdo: when you said you tried with AC3 audio, did you try playing it with digital passthrough ? |
| [23:25:36] | Beirdo: | llvm is not a proven compiler yet :) |
| [23:25:52] | Beirdo: | huh? |
| [23:25:58] | ThisNewGuy (ThisNewGuy!~doug@pool-74-102-15-78.nwrknj.fios.verizon.net) has joined #mythtv | |
| [23:26:02] | jya: | in your merge commit |
| [23:26:10] | jya: | you said you tried various audio content |
| [23:26:22] | Beirdo: | I don't have digital passthrough |
| [23:26:30] | Beirdo: | there is no way for me to test that |
| [23:26:32] | jya: | including AC3… so I'm wondering if you played an AC3 stream with digital passthrough or not |
| [23:26:37] | Beirdo: | nope |
| [23:26:49] | jya: | ok… haven't tried either yet as my mac isn't setup for it.. |
| [23:26:54] | jya: | presumably it's broken |
| [23:27:02] | Beirdo: | that is possible |
| [23:27:09] | jya: | there's a ticket about it |
| [23:27:24] | Beirdo: | there's an app for that |
| [23:27:27] | Beirdo: | oh wait. |
| [23:27:29] | jya: | :) |
| [23:27:45] | jya: | Beirdo: do you know anything much about ffmpeg and the use of "demuxer" ? |
| [23:28:01] | Beirdo: | in what context? |
| [23:28:04] | cocoa117 (cocoa117!~cocoa117@188-222-31-239.zone13.bethere.co.uk) has quit (Ping timeout: 245 seconds) | |
| [23:28:04] | ThisNewGuy1 (ThisNewGuy1!~doug@pool-74-102-15-78.nwrknj.fios.verizon.net) has quit (Ping timeout: 245 seconds) | |
| [23:28:09] | jya: | http live streaming |
| [23:28:16] | jya: | I can now play a HLS stream |
| [23:28:58] | jya: | but it gives a verbose with http://pastebin.com/M5dXf03A |
| [23:29:12] | jya: | that's when calling ffurl_open |
| [23:29:27] | Beirdo: | I think what they consider the demuxer is the container code... let me look at your spew |
| [23:29:31] | cattelan is now known as cattelan_away | |
| [23:29:47] | danielk22: | stuartm: The HDHomeRun recorder is the one recorder I've been hands off as much as possible for some time now. |
| [23:30:11] | Beirdo: | hmm, OK. how does that URL differ from the one you are playing? |
| [23:30:29] | jya: | what's i'm sure about, is that when using ffurl_open, it set up a URLProtocol |
| [23:31:07] | danielk22: | stuartm: But if you open a ticket I'll get to it eventually. |
| [23:31:09] | jya: | if I play the http://....m3u8 url, what it retrieves is the 130 bytes playlist, doesn't attempt to play it or detect that it's a HLS |
| [23:31:26] | jya: | so I play a hls+http://bha.m3u8 |
| [23:31:48] | Beirdo: | ahhh |
| [23:31:51] | jya: | this forces the hls player… |
| [23:32:11] | jya: | but this call the old applehttp |
| [23:32:18] | jya: | which is the hlsproto.c file |
| [23:32:39] | jya: | there's another HLS decoder in ffmpeg: hls.c (previously called applehttp.c) |
| [23:32:57] | jya: | that's the one I'm trying to use, as it supports much more stuff (in particular: seeking) |
| [23:33:09] | NightMonkey (NightMonkey!~NightrMon@pdpc/supporter/professional/nightmonkey) has quit (Remote host closed the connection) | |
| [23:34:16] | Beirdo: | hmmm, let me code read a moment. |
| [23:35:11] | Beirdo: | you have an example .m3u8 playlist kicking around? |
| [23:35:57] | cocoa117 (cocoa117!~cocoa117@188-222-31-239.zone13.bethere.co.uk) has joined #mythtv | |
| [23:37:31] | jya: | plenty local |
| [23:37:38] | jya: | otherwise… mythtv ? |
| [23:38:30] | jya: | backendurl:6544 -> API -> HTTP Live Streaming -> create a stream there |
| [23:38:49] | jya: | I think Apple has a HLS test stream |
| [23:38:52] | jya: | hold a sec |
| [23:40:39] | Beirdo: | it looks like they might want you to use avformat_open_input() |
| [23:41:21] | Beirdo: | rather than ffurl_open |
| [23:41:52] | Captain_Murdoch (Captain_Murdoch!~Captain_M@ctp-lxut07.bc2va.org) has joined #mythtv | |
| [23:41:52] | Captain_Murdoch (Captain_Murdoch!~Captain_M@ctp-lxut07.bc2va.org) has quit (Changing host) | |
| [23:41:53] | Captain_Murdoch (Captain_Murdoch!~Captain_M@mythtv/developer/CaptainMurdoch) has joined #mythtv | |
| [23:41:55] | Captain_Murdoch_ (Captain_Murdoch_!~Captain_M@ctp-lxut07.bc2va.org) has joined #mythtv | |
| [23:42:05] | Captain_Murdoch_ (Captain_Murdoch_!~Captain_M@ctp-lxut07.bc2va.org) has quit (Remote host closed the connection) | |
| [23:42:51] | Beirdo: | and the file probing should see the start of the .m3u8 file and know to use the hls demuxer |
| [23:43:16] | jya: | http://devimages.apple.com/iphone/samples/bip . . . g_index.m3u8 |
| [23:43:29] | jya: | cool, myth plays it |
| [23:44:05] | jya: | thanks.. let me try that |
| [23:44:14] | Beirdo: | in particular, it looks for #EXTM3U as teh first 7 bytes |
| [23:44:34] | jya: | yes, that's the m3u8 tag |
| [23:44:41] | jya: | however, it needs to check for more |
| [23:44:48] | Beirdo: | and then one of #EXT-X-STREAM-INF:, #EXT-X-TARGETDURATION: or #EXT-X-MEDIA-SEQUENCE: |
| [23:45:02] | Beirdo: | if it finds one of those, it claims it |
| [23:45:04] | jya: | you're looking at the hls.c code I assume |
| [23:45:08] | Beirdo: | yup |
| [23:45:14] | Beirdo: | hls_probe |
| [23:45:14] | jya: | is avformat_open_input a direct replacement to ffurl_open ? |
| [23:45:34] | jya: | I've trace that code line by line to see what it was doing :) |
| [23:45:43] | Beirdo: | not 100% sure |
| [23:45:50] | Beirdo: | it's documented well though :) |
| [23:46:46] | ghoti_ is now known as ghoti | |
| [23:46:51] | jya: | I'm still hesitating between using the ffmpeg hls code that is over 1 year old, and has seen lots of debug... |
| [23:47:29] | jya: | or use mine, that I've almost completed (didn't know about the ffmpeg one) but it handles things like bandwidth adaptation and changing bitstream on the fly |
| [23:48:06] | jya: | the ffmpeg's hls developer said that it's up to the application changing stream, however, they do not provide any data to allow making the decision on which stream to use |
| [23:48:09] | Beirdo: | http://ffmpeg.org/doxygen/trunk/group__lavf__ . . . tml#_details |
| [23:48:36] | jya: | could always patch their hls.c code and that's one more patch for you to worry about next time :) |
| [23:49:01] | Beirdo: | The packet data belongs to the demuxer and is invalid after the next call to av_read_frame(). |
| [23:49:07] | Beirdo: | Damn, would you look at that :) |
| [23:49:34] | Beirdo: | thanks for fixing that, although it makes it look like pointless memcpy, but it's not pointless at all |
| [23:50:29] | Beirdo: | well, if we patch their code, we could consider handing it back to them and trying to get it incorporated |
| [23:50:37] | Beirdo: | s/could/should/ really |
| [23:50:38] | jya: | I wonder why we bother doing our own filebuffer.cpp and demux files and mpeg ourselves… looks to me that a single ringbuffer for both local files and network (non-myth) could be handled entirely by ffmpeg avformat_open_input |
| [23:51:05] | Beirdo: | it's our protocol handling of myth:// I think |
| [23:51:16] | jya: | Beirdo: yes, there's two part to it |
| [23:51:16] | Beirdo: | but we do use that methodology |
| [23:51:24] | jya: | one is myth:// ; the other is local file |
| [23:51:29] | Beirdo: | true |
| [23:51:37] | jya: | code is rather complex, determining PTS and all |
| [23:51:45] | jya: | when it's all in ffmpeg for everyone to use |
| [23:52:31] | Beirdo: | we use avformat_open_input, along with our own AVIOContext for the ringbuffer |
| [23:52:45] | Beirdo: | it was a royal pain to debug when the API on that all changed around :) |
| [23:54:41] | dissent (dissent!~Ponzo@ponzo.ponzo.net) has quit (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.) | |
| [23:54:55] | vallor (vallor!~Ponzo@pdpc/supporter/monthlygold/vallor) has joined #mythtv | |
| [23:55:38] | stichnot (stichnot!chatzilla@mythtv/developer/stichnot) has quit (Ping timeout: 240 seconds) | |
| [23:57:26] | kormoc (kormoc!~kormoc@mythtv/developer/kormoc) has quit (Quit: kormoc) | |
IRC Logs collected by
BeirdoBot.
Please use the above link to report any bugs.